Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Markov property of the trajectory law across a gap of hhh steps

Proved
TPG.integral_head_mul_gap_tail

by Wenqian Xing · Oct 12, 2026 · Mathlib 0df444a (Lean v4.33.1)

causal-inferencemarkov-chainnonstationaritypolicy-gradienttreatment-effect

Setting. SSS is a finite state space, actions are Zt∈{0,1}Z_t\in\{0,1\}Zt​∈{0,1}, time is t=0,1,…,T−1t=0,1,\dots,T-1t=0,1,…,T−1 (the source's time t+1t+1t+1). The environment has transition matrices PztP_z^tPzt​, initial distribution ρ\rhoρ and reward laws R(⋅∣x,z)R(\cdot\mid x,z)R(⋅∣x,z) on R\mathbb RR. For a sequence q=(qt)t≥0q=(q_t)_{t\ge0}q=(qt​)t≥0​ of treatment probabilities, Pq\mathbb P_qPq​ and Eq\mathbb E_qEq​ denote the law of the trajectory (Xt,Zt,Yt)t<T(X_t,Z_t,Y_t)_{t<T}(Xt​,Zt​,Yt​)t<T​ when X0∼ρX_0\sim\rhoX0​∼ρ, the actions are independent with P(Zt=1)=qt\mathbb P(Z_t=1)=q_tP(Zt​=1)=qt​, Xt+1∼PZtt(Xt,⋅)X_{t+1}\sim P^t_{Z_t}(X_t,\cdot)Xt+1​∼PZt​t​(Xt​,⋅), and the rewards are conditionally independent with Yt∼R(⋅∣Xt,Zt)Y_t\sim R(\cdot\mid X_t,Z_t)Yt​∼R(⋅∣Xt​,Zt​) (Env.trajLaw T q); the Bernoulli policy πθ\pi_\thetaπθ​ is qt≡θq_t\equiv\thetaqt​≡θ.

Consider the horizon m+1+h+nm+1+h+nm+1+h+n. The head is the first m+1m+1m+1 time steps 0,…,m0,\dots,m0,…,m; after a gap of hhh steps, the tail is the last nnn steps m+1+h,…,m+h+nm+1+h,\dots,m+h+nm+1+h,…,m+h+n, re-indexed from 000. Let Π=Pqm+1m+1Pqm+2m+2⋯Pqm+hm+h\Pi=P^{m+1}_{q_{m+1}}P^{m+2}_{q_{m+2}}\cdots P^{m+h}_{q_{m+h}}Π=Pqm+1​m+1​Pqm+2​m+2​⋯Pqm+h​m+h​ with Pθs=θP1s+(1−θ)P0sP^s_\theta=\theta P^s_1+(1-\theta)P^s_0Pθs​=θP1s​+(1−θ)P0s​ (the identity when h=0h=0h=0), and for a state bbb let E(m+1+h,b)\mathbb E^{(m+1+h,b)}E(m+1+h,b) denote expectation for the process restarted at time m+1+hm+1+hm+1+h from bbb.

Lemma. Let 0≤qt≤10\le q_t\le10≤qt​≤1 for all ttt, and let FFF (a function of a trajectory of length m+1m+1m+1) and GGG (a function of a trajectory of length nnn) be bounded and measurable. Then

Eq[F(head) G(tail)]=∑a∈S∑b∈S Eq[F⋅PZmm(Xm,a)] Π(a,b) E(m+1+h,b)[G].\mathbb E_q\bigl[F(\text{head})\,G(\text{tail})\bigr]=\sum_{a\in S}\sum_{b\in S}\ \mathbb E_q\bigl[F\cdot P^{m}_{Z_m}(X_m,a)\bigr]\ \Pi(a,b)\ \mathbb E^{(m+1+h,b)}[G].Eq​[F(head)G(tail)]=a∈S∑​b∈S∑​ Eq​[F⋅PZm​m​(Xm​,a)] Π(a,b) E(m+1+h,b)[G].

The head influences the tail only through the distribution of the state at time m+1m+1m+1, propagated by the hhh intermediate kernels. Combined with the contraction of Assumption 1 this yields geometric decay of covariances between functions of time windows separated by hhh steps, which is how the source obtains its variance bound and the strong mixing property.

Preamble
import Mathlib
import Definitions.Def_TPG_Model
import Definitions.Def_TPG_Estimators
import Definitions.Def_TPG_Markov
import Definitions.Def_TPG_Bridge

open MeasureTheory ProbabilityTheory
open scoped BigOperators ENNReal
Formal statement
namespace TPG
theorem integral_head_mul_gap_tail {S : Type} [Fintype S] [DecidableEq S] [MeasurableSpace S] [MeasurableSingletonClass S]
    (E : Env S) (m h n : ℕ) (q : ℕ → ℝ) (hq : ∀ t : ℕ, 0 ≤ q t ∧ q t ≤ 1)
    (F : Traj S (m + 1) → ℝ) (G : Traj S n → ℝ) (hF : Measurable F) (hG : Measurable G)
    (CF CG : ℝ) (hFb : ∀ ω, |F ω| ≤ CF) (hGb : ∀ ω, |G ω| ≤ CG) :
    ∫ ω, F (headTraj (m + 1) (h + n) ω) * G (tailTraj h n (tailTraj (m + 1) (h + n) ω))
        ∂(E.trajLaw (m + 1 + (h + n)) q)
      = ∑ a : S, ∑ b : S,
          (∫ ω, F ω * E.P m (Z (Fin.last m) ω) (X (Fin.last m) ω) a ∂(E.trajLaw (m + 1) q))
            * (((List.Ico (m + 1) (m + 1 + h)).map (fun s => E.PMix (q s) s)).prod a b)
            * ∫ ω, G ω ∂((E.shift (m + 1 + h) b).trajLaw n (fun t => q (m + 1 + h + t))) := by sorry
end TPG
Source
Johari, Peng, Xing, Estimation of Treatment Effects Under Nonstationarity via the Truncated Policy Gradient Estimator, revised manuscript of October 2026 (revision of arXiv:2506.05308v3, https://arxiv.org/abs/2506.05308); numbering follows the revised manuscript. Auxiliary statement, not a numbered result of the paper: it formalizes the identity E[B_j | F_i] = (P_{1/2}^{(i+1)→a_j} g_j)(X_{i+1}) of Appendix F.1, p. 40, and the conditioning step in the proof of Lemma 2, p. 28.

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me